PredictionBook is now read-only ( read more ).

Voice recognition software will be capable of replacing the average court stenographer in five years.

Created by lavalamp on 2013-02-02; known on 2018-02-01; judged wrong by lavalamp on 2018-02-01.

  • lavalamp estimated 45% on 2013-02-02
  • gwern said “they obviously won’t actually replace them, so how are you judging ‘will be capable’?” on 2013-02-03
  • lavalamp said “I have experience and friends in the field. I’ll look at what’s available at the time. If it seems like the only obstacle is slowness to change in the industry (regulatory or otherwise), I’ll judge this true, otherwise false.” on 2013-02-03
  • lavalamp said “Specifically, mask reporting does not count, nor will any form of reporting where a user repeats what is said to the VR software.” on 2013-02-03
  • Roxolan estimated 33% on 2013-02-03
  • Tuxedage estimated 20% on 2013-02-06
  • moridinamael estimated 60% on 2013-02-06
  • simplicio estimated 55% on 2013-02-09
  • DaFranker estimated 30% on 2013-02-11
  • halfwaytogallifrey estimated 1% and said “Court room acoustics and dialects with zero ability to train the software for each speaker would make this impossible even if voice recognition somehow edged far closer to 100% accuracy in a controlled environment.” on 2013-02-11
  • DSanders estimated 10% and said “Computing power is not the problem. Voice recognition requires a larger database and more ‘training’. Even so, people would emotionally rather rely on stenographers even if VR quality was better.” on 2013-02-14
  • Houshalter estimated 25% on 2013-02-15
  • themusicgod1 estimated 40% and said “imho court room acoustics is already not a problem. dialects …depends who’s serious about solving the problem. Google is getting good at dialects when they want to be. Multiple voices might be issue” on 2016-07-10
  • lavalamp said “http://www.martelelectronics.com/worlds-first-multiple-voice-recognition-recorder-conferences-meetings-automatic-voice-to-text-system/” on 2018-02-01
  • lavalamp said “https://www.ibm.com/blogs/watson/2016/12/look-whos-talking-ibm-debuts-watson-speech-text-speaker-diarization-beta/” on 2018-02-01
  • lavalamp said “Looks like this got very close. One more year and I’d probably have to mark this right.” on 2018-02-01
  • lavalamp   judged this prediction wrong on 2018-02-01.